ClawHub
Back to skill

Security audit

Dlazy Image Storyboard

Security checks across malware telemetry and agentic risk

Overview

This storyboard skill uses a disclosed dLazy cloud CLI to generate images after user confirmation, with expected API key storage and remote prompt/media processing.

Install only if you are comfortable sending storyboard prompts, parameters, and any referenced media files to dLazy's cloud service. Use npx if you do not want a persistent global CLI, avoid putting sensitive content in prompts or uploaded files, and rotate or revoke the stored dLazy API key if needed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The skill materially expands from storyboard planning into instructing the agent to execute terminal commands that call an external image-generation SaaS. That creates a real security boundary crossing: user prompts and local file paths may be sent off-host, and command execution is being normalized from within documentation. In this context, the issue is more dangerous because the skill explicitly frames the agent as able to run commands and prescribes operational behavior, increasing the chance of unintended data disclosure or unsafe tool use.

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The documentation states the skill itself will not access the network or file system, yet the required workflow relies on a CLI that uploads local files to remote services and retrieves hosted results. This misleading assurance can cause operators to approve the skill under false assumptions about data handling, leading to unintentional exfiltration of prompts or local media. The contradiction makes the skill context more dangerous because users may trust the safety claim and skip scrutiny of the external data flows.

Missing User Warnings

Low
Confidence
84% confidence
Finding
The execution instructions tell the agent to send prompts and possibly referenced media through the dlazy CLI, but the step-by-step generation section does not repeat or foreground that user content will be transmitted to third-party services. In a skill that actively instructs terminal execution and cloud rendering, this omission can mislead users into sharing sensitive text or local media without sufficiently informed consent, increasing privacy and data-handling risk.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.